Mary Grant, a marketing manager of an online retailer, has built two predictive models for estimating the annual spending of new customers. Applying the models to the 100 observations in the validation data set generates a table that lists the customers’ actual spending and predicted spending. A portion of the data set is shown below.

a. Compute the ME, RMSE, MAD, MPE, and MAPE for the two predictive models. 

b. Are the predictive models over- or underestimating the actual selling price on average? 

c. Compare the RMSE and MAD of the two predictive models. Which predictive model performs better? 

d. Compare the better-predictive model to a base model where every customer is predicted to have the average spending of the cases in the training data set, which is $290. Does the better-predictive model built by Mary outperform the base model in terms of RMSE?

Transcribed Image Text:

Customer 1 2 1 100 Actual Spending 555 190 ⠀ 464 Predicted Spending 1 587 198 : 380 Predicted Spending 2 487 182 T 312